[lld-macho] Change UUID calculation to use constant chunk count

[keith]

Summary

Using the current machine's thread count as the chunkCount results in
non-deterministic binaries across machines with different core counts.
This is an attempt to pick a relatively reasoanble number for the
average case. Theoretically this could be configurable.

Fixes: https://github.com/llvm/llvm-project/issues/63961

[int3]

seems reasonable

[MaskRay]

I think it may be better to use a fixed chunk size instead of fixed chunk count :)

[keith]

I'm happy for us to just go with your patch if you think that's preferred!

[MaskRay]

In D155733#4516740, @keith wrote:

I'm happy for us to just go with your patch if you think that's preferred!

Thank you! I reclaimed D155761.

(I should have read the linked bug first. I just saw your other patch and thought that the uuid thing can be changed based on my understanding of ld.lld --build-id :))

[keith]

https://reviews.llvm.org/D155761

lld/MachO/Writer.cpp

 
 // In order to utilize multiple cores, we first split the buffer into chunks,
 // compute a hash for each chunk, and then compute a hash value of the hash
 // values.
 void Writer::writeUuid() {
   TimeTraceScope timeScope("Computing UUID");
 
   ArrayRef<uint8_t> data{buffer->getBufferStart(), buffer->getBufferEnd()};
-  unsigned chunkCount = parallel::strategy.compute_thread_count() * 10;
+  unsigned chunkCount = 100;
   // Round-up integer division
   size_t chunkSize = (data.size() + chunkCount - 1) / chunkCount;
   std::vector<ArrayRef<uint8_t>> chunks = split(data, chunkSize);
   // Leave one slot for filename
   std::vector<uint64_t> hashes(chunks.size() + 1);
   SmallVector<std::shared_future<void>> threadFutures;
   threadFutures.reserve(chunks.size());
   for (size_t i = 0; i < chunks.size(); ++i)